Point of Care testing, critical care, emergency medicine ...

W140

Point of Care testing, critical care, emergency medicine.

HIGH-SENSITIVITY TROPONIN ON ATELLICA VTLI PATIENT-SIDE IMMUNOASSAY ANALYZER (SIEMENS): EVALUATION,RISK ASSESSMENT AND PRACTICABILITY

B. Beneteau-Burnat 1, C. Hermand 2, P. Pernet 1, F. Bonnet 2, M. Etienne 1, L. Terracol 1, A. Suparschi 1, C. Aussel 1, S.

Ruelle 1, M. Vaubourdolle 11Biochemistry Laboratory, Department of Biochemistry-Hormonology, Saint-Antoine Hospital, DMU BioGeM, AP-HPSorbonne University, Paris, France2Emergency Department, Saint-Antoine Hospital, DMU DEMIURGE, AP-HP Sorbonne University, Paris, France

BACKGROUND-AIM

To evaluate the Atellica VTLI analyzer (Siemens), a new handheld device for high-sensitivity (HS) troponin I testing atpoint-of-care (POC), producing results in 8 minutes from a small sample volume (30-100 µL) of whole blood, capillaryor venous samples. The system consists of a small instrument (<1 kg) with disposable cartridges and is based on theprecisely controlled motion of magnetic beads coated with antibodies. Evaluation was performed in the laboratory andin the Emergency Department (ED) with POC users.

METHODS

Evaluation was performed according to the French Society of Clinical Biology (SFBC): within-run precision and between-day precision at 3 levels, comparison of methods with 110 routine patient samples, reportable range. Comparisonanalyzer was our biochemistry routine Atellica IM (Siemens) with a HS troponin I assay (dual capture sandwichimmunoassay). Risk assessment was done using Ishikawa diagram and FMECA method. Practicability was evaluatedusing a survey (17 questions) in the laboratory by 15 users and 15 POC users in the ED.

RESULTS

Within-run and between-day imprecision gave conform CV values. Linearity was within the expected range given bySiemens (from limit of quantification up to 1250 ng/L). Method comparisons showed a close agreement (r = 0.96) butresults with the Atellica VTLI assay were 3 times lower than those of the Atellica IM assay (y = 0.33x + 6.7 ng/L). Riskassessment showed Atellica VTLI very well designed for a POC or a stat laboratory usage. Users, both in the laboratoryand the ED, gave positive reviews and believe there is a strong benefit for patients to have this device in POC.

CONCLUSIONS

Atellica VTLI analytical performances for POC HS troponin are validated with the reliability of traditional systems. Thisdevice fulfils all the requirements for a POC or a stat laboratory usage. However, hospitals equipped with Atellica IM intheir routine labs will have to take into account for patient’s follow-up that methods and results are different althoughbelonging to the same “Atellica” brand from Siemens. A strong educational program for clinicians will be essential.

Poster Abstracts – EuroMedLab Munich 2021 – Munich, Nov 28-Dec 02, 2021 • DOI 10.1515/cclm-2021-5033 Clin Chem Lab Med 2021; 59, Special Suppl, pp S94 – S835, Nov/Dec 2021 • Copyright © by Walter de Gruyter • Berlin • Boston S705

W141


POINT-OF-CARE ASSESSMENT OF BIOCHEMISTRY PANELS: COMPARISON OF TWO MULTI-PURPOSE DEVICES

L. Blairon 1, R. Cupaiolo 1, A. Wilmet 1, I. Beukinga 1, M. Tré-Hardy 11Laboratory Medicine, Iris Hospitals South, Brussels

BACKGROUND-AIM

POCT instruments make it possible to measure more and more biological parameters at the patient's bedsideand facilitate rapid diagnosis. However, the number of parameters per device is often limited. As a consequence,multiplying the number of POCT analyses often means increasing the number of instruments. Apart from bloodgas analyzers, which are generally large, there are few instruments able to carry out several determinations at thesame time. In this study, we evaluated Piccolo Xpress and LINX EVO, 2 compact analyzers capable of measuring tenbiochemical parameters at once.

METHODS

Routine serum samples previously assessed on a cobas® 8000 platform (Roche Diagnostics, Mannheim, DE) weretested with the ER III and Biochemistry Plus panels respectively on LINX EVO (A.Menarini Diagnostics, Firenze, IT)and Piccolo Xpress (Abaxis, Inc., Union City, CA, USA). The slope, intercept and coefficient of determination R² werecalculated for each available parameter. Values outside the lower and upper measurement range of both LINX EVO andPiccolo Xpress were removed from the statistical analyses.

RESULTS

49 clinical samples were analyzed in parallel with ER III and Biochemestry Plus panels. All the coefficients ofdeterminations after comparison to the cobas® were >0.90, with the exception of sodium on LINX EVO (0.87) and ofcalcium (0.83) and albumin (0.88) on Piccolo Xpress. All regression equations were excellent, with slopes between 0.85and 1.13, except for albumin (0.78) and amylase (1.29) on Piccolo Xpress.

CONCLUSIONS

To our knowledge, LINX EVO and Piccolo Xpress are the only POCT devices which are able to assess several biochemistryanalyses at the same time in approximately 15 minutes. Their degrees of agreement are very good compared to ourroutine biochemistry system. Both panels are suitable for use in emergency medicine, however ER III provides resultsfor total bilirubin and lipase, two important parameters for the management of abdominal pain. Using such integratedsolution in emergency rooms will help the physician in a faster management and throughput of patients.


W142


TURNING A LABORATORY HAEMATOLOGY INSTRUMENT INTO A POINT OF CARE TESTING (POCT)

N. Modi Omasombo 2, D. Izgarevic 1, S. Collignon 1, R. Cupaiolo 2, M.d.C. Hernando Axpe 2, I. Beukinga 2, M. Tré-Hardy2, F. Coenen 1, L. Blairon 21Emergency Medicine, Iris Hospitals South, Brussels2Laboratory Medicine, Iris Hospitals South, Brussels

BACKGROUND-AIM

Although there are a multitude of POCT instruments for the determination of chemistry parameters, there are fewfor haematology. Also we did not have the opportunity to test the pocH-100i (Sysmex, Kobe, Japan) for reasons ofunavailability in Belgium. We installed a Beckman DxH-520 (Beckman-Coulter, Brea, CA, USA) laboratory analyser inan emergency room (ER) so that it could be used as a point-of-care instrument in order to objectify the time saved bytesting the patient directly in the ER.

METHODS

For one week, all patients requiring a blood sample for a haematology formula were drawn with an additional EDTAtube for in situ analysis. The clinician was asked to record the time of the blood test to assess the difference in turn-around time (TAT) between the instrument used as a POCT and the central laboratory analysis.

RESULTS

Forty consecutive patients were included: median age 59.4 years (95% CI: 53.2-65.6); sex ratio M/F=1/2.1. For 2patients, DxH-520 did not provide a leukocyte formula due to probable handling errors by emergency users. The medianTAT (95% CI) for the DxH-520 and in situ analysis were 10 min (6-14) and 32 min (16.9-37.1) respectively (P<0.0001,Mann-Whitney U-test).

CONCLUSIONS

The DxH-520, although not a POCT, was used as such and helped speed up the rendering of haematology resultssignificantly. However, even if its use is simple for a laboratory professional, the interface is neither simple nor intuitivefor nurses and physicians. In addition it has external reagents and waste, and required a lot of maintenance. However,it has advantages such as speed and small size and could be adapted for use like a POCT: prevention of errors inuse, POCT-1A connectivity, simplified display and unique reagent/waste cassette like POCT blood gas analysers. Wetherefore encourage the manufacturer to develop a solution in the way of POCT. The availability of such an instrumentin an oncology day hospital could bring a real added-value because the initiation of chemotherapy depends on therapid achievement of the complete blood count.


W143


EVALUATION OF TWO POINT-OF-CARE INSTRUMENTS FOR LIPID AND GLUCOSE DETERMINATION

N. Modi Omasombo 1, R. Cupaiolo 1, A. Wilmet 1, I. Beukinga 1, L. Blairon 11Laboratory Medicine, Iris Hospitals South, Brussels

BACKGROUND-AIM

POCT instruments are primarily oriented for the determination of vital or urgent parameters. However, decentralizedmeasurement tools are also used in consultation rooms, including for monitoring chronic diseases such as diabetesor cardiovascular diseases. Some authors have also studied the impact of an immediate outcome on the patient’scompliance with the regimen and medication. In this study, we evaluated the degree of agreement of the lipid panelsof LINX DUO and LINX EVO, two POCT instruments that can be used for both emergency and chronic disease follow-uptesting by comparing them to our routine biochemistry technique.

METHODS

57 samples of serum from the routine tested on cobas 8000 (Roche Diagnostic, Mannheim, DE) with a wide variation incholesterol, triglyceride and HDL-cholesterol values were selected for POCT analysis. The Lipid Panel Plus was used onLINX DUO and the Lipid Panel on LINX EVO (A.Menarini Diagnostics, Firenze, IT) for the measurement of triglycerides,total cholesterol, HDL and glucose. LDL and VLDL are calculated parameters. Values outside the measurement limitswere removed from the statistical analysis. The glucose results were compared to the performance criteria defined ininternational standards ISO 15197, POCT 12-A3 and FDA Draft Guidance Blood Glucose Monitoring Test Systems forPrescription Point-of-Care Use.

RESULTS

57 samples were tested with LINX EVO and 54 with LINX DUO. The coefficients of determination (R²) were ≥0.89 for allthe lipids, with a slope between 0.88 and 1.11 and between 0.84 and 1.18 for LINX DUO and LINX EVO respectively. Forglucose, R² were ≥0.96 with a slope of 0.97 for LINX DUO and of 1.00 for LINX EVO. The concordance of LINX EVO withISO 15197, POCT 12-A3 and FDA criteria were 100%, ≥96% and ≥95.9% respectively. The concordance of LINX DUO withISO 15197, POCT 12-A3 and FDA criteria were ≥92%, ≥88% and ≥93.6% respectively.

CONCLUSIONS

LINX DUO and LINX EVO are two decentralized biology instruments that are suitable for lipid dosing in a preventivemedicine or nutritionist’s office. In addition, LINX EVO shows outstanding performance in glucose determination. Theturnaround time of the LINX EVO is 15 minutes for one sample versus 8 minutes for 1 or 2 samples at a time on LINXDUO.


W144


INFLUENCE OF A POCT MULTI-PARAMETER BIOCHEMISTRY ANALYZER ON THE MANAGEMENT OF PATIENTS WITHABDOMINAL PAIN IN EMERGENCY DEPARTMENT

S. Collignon 1, P. Djimafo Tiwet 1, M. Abbasi 1, R. Cupaiolo 2, M. Tré-Hardy 2, I. Beukinga 2, F. Coenen 1, L. Blairon 21Emergency Medicine, Iris Hospitals South, Brussels2Laboratory Medicine, Iris Hospitals South, Brussels

BACKGROUND-AIM

Abdominal pain (AP) accounts for 5 to 10% of the admissions to emergency departments (ED). Management of APsusually requires clinical biology tests and imaging. Except in major emergencies, clinicians wait for the labs resultsto refer the patient to a CT-scanner (contrast-enhanced or not), or to an ultrasound. We evaluated the time saved inthe management of ED-patients with AP by analyzing blood samples in parallel on a POCT device performing severalbiochemistry tests in 15 minutes.

METHODS

An additional Lithium-heparin tube was drawn for the POCT tests. POCT analyses were performed with the ER III panelon LINX EVO (A.Menarini Diagnostics, Firenze, IT). The usual blood sample was sent at the same time to the centrallaboratory. The response time for biochemistry analyses was recorded according to the LINX EVO timer. The responsetime of the laboratory was recorded regarding the log of the lab information system. The time frames were calculatedbased on the time of blood collection.

RESULTS

62 patients were included (median age [95% CI]: 46.6 [41.5-51.6]). The median response times to the patient recordwere 21 [16.6-25.4] minutes for LINX EVO and 90 [50.2-129.8] minutes for the routine blood analyses. The mediantime gain in the availability of results on LINX EVO was 58 [17.5-98.5] minutes (p<0.001). The imaging examinationsperformed were CT with (N=30) and without contrast enhancement (N=12), ultrasound (N=10), plain abdomen film(N=2), none (N=10). 25 patients were hospitalized, 2 were temporary hospitalized and 35 were discharged. Finaldiagnosis were common AP (N=6), biliary pain (N=8), gastric ulcer (N=5), gastroenteritis (N=6), gynecological pain(N=3), obstruction (N=5), pancreatitis (N=2), peritonitis/perforation/appendicitis (N=7), urolithiasis (N=3), urinary tractinfection (N=4), other (N=13).

CONCLUSIONS

LINX EVO allowed a median gain of 58 minutes in patient management: need and choice of an imaging examination,possibility of contrast agent injection, diagnosis of inflammatory syndrome. The ER III panel was previously evaluatedin our laboratory and showed performance equivalent to routine instruments. To our knowledge, it is the only POCTdevice able to dose lipase which is essential for the management of APs.


W145


EVALUATION OF THE EMERGENCY PANEL ER III ON THE LINX EVO ANALYZER


BACKGROUND-AIM

The LINX EVO analyzer (A.Menarini Diagnostics, Firenze, IT) is a new POCT instrument that enables the dosage ofseveral biochemistry parameters in 15 minutes in whole blood or serum. The manufacturer offers 11 different reagentdiscs which allow dosing combinations of 4 to 13 parameters. Some additional parameters are calculated. We aimed toevaluate the ER III panel, which is to be the most suitable one for rapid dosages in emergency rooms. The parameterstested are: C reactive protein, glucose, aspartate and alanine transaminases, gamma-glutamyltransferase, bilirubin,amylase, lipase, blood urea nitrogen, creatinine, sodium and potassium.

METHODS

Intra-assay precision was calculated by testing 20 times 2 levels of Liquid Unassayed Multiqual® controls (Bio-Rad,Hercules, CA). Inter-assay precision was calculated by testing the same controls for 20 consecutive working days.Matrix correlation was evaluated on 10 clinical samples collected both on serum and lithium-heparin tubes. The degreeof agreement was assessed by comparing the results from the ER III to 73 clinical samples analyzed on a cobas®8000 (Roche, Mannheim, DE). The clinical samples were chosen to obtain a range of low to high values within themeasurement range of LINX EVO.

RESULTS

Coefficients of variation of ER III ranged from 0% to 5.5% for intra-assay precision and from 0% to 5.8% for inter-assay,excepted for low bilirubin values (10.7%). The slopes ranged from 0.86 to 1.14. All coefficients of determination (R²)were >0.90, excepted for sodium (0.86). The matrix correlation showed R² >0.93 and slopes between 0.962 and 1.044for all parameters but bilirubin, sodium and potassium.

CONCLUSIONS

The ER III panel allows to measure 12 biochemistry parameters simultaneously (plus one calculated) in 15 minutes.Even if some parameters (ions, glucose, and bilirubin) can be measured on a blood gas analyzer, often available in theemergency rooms, the panel provides other valuable results for rapid management of patients in the emergency room.The results are very reliable compared to our large platform of routine biochemistry. The artificially poor R² valuesof the matrix comparison for Na, K and bilirubin are due to the small number of samples combined to the limiteddispersion of the values.


W146


QUANTITATIVE MEASUREMENT OF C-REACTIVE PROTEIN (CRP) BY POCT WITH LINX DUO ANALYZER AND COMPARISONTO TWO OTHER INSTRUMENTS


BACKGROUND-AIM

Respiratory tract infections (RTIs) are the most frequent infections in primary care (about 23% of general practice). Wedistinguish the upper (pharyngitis, tonsillitis, laryngitis, rhinosinusitis, otitis media) and lower (pneumonia, bronchitis,tracheitis, exacerbation of COPD) RTIs. Most RTIs are of viral origin, but a small number are due to bacteria and canthus respond to antibiotics (AB). However, overprescription of AB is common in primary care with risk of antimicrobialresistance. CRP testing can be an interesting tool to confirm the bacterial cause of RTIs in general practice. Studieson POCT CRP seem to conclude that CRP is better at ruling in than in ruling out bacterial cause of RTIs. In this study,we tested LINX DUO, a POCT analyzer recently available on the market and compared it to 2 other devices of the sametype marketed longer.

METHODS

A panel of samples was selected from fresh samples tested on cobas 8000 (Roche Diagnostics, Mannheim, DE),used as a comparison method. Samples were selected to achieve a dispersion of results between low and very highvalues. Samples were tested fresh and under the same conditions with LINX DUO (A.Menarini Diagnostics, Firenze,IT), Cube-S (Eurolyser Diagnostica, Salzburg, AT) and Afinion 2 (Abbott Diagnostics, Lake Forest, IL, USA) according tomanufacturers' instructions. Values obtained outside the measurement range of the devices have been removed fromthe statistical analysis (Afinion 5-160 mg/dL; Cube-S 1-180 mg/dL; LINX DUO 5-200 mg/dL).

RESULTS

The number of samples tested with Afinion 2, Cube-S and LINX DUO were 89, 92 and 86 respectively. The respectiveequations of the regression curves were y=1.06x-0.92 (r²=0.99); y=0.94x+1.63 (r²=0.95); y=1.22x-0.13 (r²=0.92). With acutoff of 10 mg/dL for the suspicion of the bacterial etiology of pneumonia, the respective sensitivity/specificity [95%CI]; Cohen’s kappa (κ) score were 97.4% [91%-99%]/84.6% [58%-96%]; κ=0.82; 94.7% [87%-98%]/94.1% [73%-99%];κ=0.83; 97.3% [91%-99%]/90.9% [62%-98%]; κ=0.85.

CONCLUSIONS

LINX DUO is a reliable alternative to the other POCT methods for the quantitative measurement of CRP. Its performanceis comparable to that of the other 2 instruments tested.


W147


EVALUATION OF THE DOSAGE OF C-REACTIVE PROTEIN WITH LINX DUO


BACKGROUND-AIM

LINX DUO (A.Menarini Diagnostics, Firenze, IT) is a point-of-care testing instrument that is able to assess the dosageof C-reactive protein (CRP), lipids, and glycated hemoglobin. In this study, we aimed to evaluate the precision and theaccuracy of the CRP reagent kit.

METHODS

Intra-assay precision was performed using two clinical samples on EDTA, with low (11.3 mg/dL) and high (47.8 mg/dL)CRP values preliminary tested on a cobas 8000 chemistry module c 702 (Roche Diagnostic, Mannheim, DE). Those twosamples were tested for 20 consecutive runs. Inter-assay precision was performed using Randox Acusera Liquid CRPControl levels 2 and 3 (Randox Laboratories LTD, Crumlin, UK) for 20 days. The degree of agreement was assessed bytesting 80 serum samples from the routine tested with the cobas c 702 module. All the samples were included into therange of measurement of the LINX DUO (5-200 mg/dL).

RESULTS

Intra-assay precision was 18.3% for low values and 8.9% for high values. Inter-assay precision was 8.5% for low valuesand 8.3% for high values. The regression curve obtained after comparison to the cobas c 702 module was y = 1.20x -0.38 with a coefficient of determination of 0.96. The mean bias was -14.2%. The total errors for low and high valueswere 20.9% and 2.5% respectively (95% confidence interval). The sigma scores for low and high values were 4.3 and8.7 respectively.

CONCLUSIONS

The CRP kit on LINX DUO is a precise and accurate tool for the measurement of CRP at the bedside. It allows aquantitative dosage of CRP between 5 and 200 mg/dL in about 6 minutes. The total errors observed for both low andhigh values are much lower than the allowable total error.


W148


LOW PARTIAL PRESSURE OF OXYGEN CAUSES SIGNIFICANT UNDERESTIMATION OF GLUCOSE CONCENTRATIONMEASURED BY GEM 5000 BLOOD GAS ANALYZER IN WHOLE BLOOD

Y. Butorin 1, A. Venner 1, L. De Koning 1, H. Paul 1, I. Seiden Long 1, H. Sadrzadeh 11Alberta Precision Laboratories

BACKGROUND-AIM

Glucose determination by GEM blood gas analyzer manufactured by Instrumentation Laboratory (IL) is based onenzymatic reaction of glucose with oxygen in the presence of glucose oxidase and the electrochemical oxidation ofthe resulting hydrogen peroxide at the platinum electrode. It was noted that measurement of glucose on GEM 4000,was subject to significant negative bias in the presence of low partial pressure of oxygen (pO2) in whole blood sample.The aim of this study was to verify if new GEM 5000 analyzer is underestimating concentration of glucose in wholeblood in the presence of low pO2.

METHODS

Whole blood samples from healthy volunteers have been collected in BD Li Heparin vacutainers and artificially depletedof oxygen after sitting at room temperature for 72 hours. Subsequently, samples have been spiked with glucosesolution and serial measurement of blood gas panel was performed on GEM 5000 blood gas analyzer, with oxygencontent gradually increasing. Glucose concentration in the sample was confirmed using Siemens Vista chemistryanalyzer. The difference between serial glucose measurements obtained on the blood gas analyzer and confirmatorytesting on Siemens Vista was plotted against the pO2 content of the whole blood sample.

RESULTS

Concentration of glucose measured by GEM 5000 was up to 29% lower than that on Siemens Vista analyzer at pO2 of25 mmHg. The rate of glucose underestimation by GEM 5000 was decreasing as pO2 was increasing. The differencebetween two methods becomes clinically insignificant at pO2 greater than 30-35 mmHg, depending on the lot of GEMcartridges.

CONCLUSIONS

Whole blood glucose measurement by GEM 5000 blood gas analyzer is subject to significant negative interference inthe presence of pO2 below 35 mmHg.


W149


ASSESSMENT OF THE ANALYTICAL PERFORMANCES OF STATSTRIP GLU/KET FOR GLUCOSE MEASUREMENT INCEREBROSPINAL FLUID

R. Cupaiolo 1, M. Tré-Hardy 1, A. Wilmet 1, I. Beukinga 1, L. Blairon 11Laboratory Medicine, Iris Hospitals South, Brussels

BACKGROUND-AIM

Measurement of cerebrospinal fluid (CSF) glucose (GLU) provides essential information to the clinician:hypoglycorrhachia is a symptom of various abnormal clinical conditions and is also a useful piece of evidence in thedifferential diagnosis of bacterial (low GLU) and viral (normal GLU) meningitis. GLU biochemistry assays require samplecentrifugation which may delay time-to-result. In this study, we evaluated the measurement of GLU in CSF using ablood GLU monitoring systems (BGMS). Our goal is to propose this method to clinicians to obtain a rapid GLU result inCSF with a minimal sample volume requirement of 0.6µl at the bedside.

METHODS

We submitted a routine StatStrip GLU/KET meter (Nova Biomedical, Waltham, USA) to an intra-assay (10 runs with 2 lotsof strips) and inter-assay (twice daily for 20 days in duplicate with 2 lots of strips) precision using clinical samples of CSFwith low and high GLU level. Method comparison was performed using consecutive clinical CSF samples and comparingresults obtained on the biochemistry routine platform (cobas 8000, Roche Diagnostics, Mannheim, DE) versus StatStripGLU/KET (duplicate with 2 lots of strips).

RESULTS

The coefficients of variation observed for the intra-assay were 6.2% on average for both the low (18 mg/dL) and high(58 mg/dL) value samples. The CVs for inter-assay (median (95% CI)) were 9.6% (9.3-13.0) and 8.0% (6.4-8.6) for the low(18 mg/dL) and high (75 mg/dL) levels respectively. The analysis of the precision on 60 clinical samples (median age;95% CI: 26.7; 0.2-48.4) showed a slope of 0.81 and a coefficient of determination (R²) of 0.94, with an underestimationof the high values (>100 mg/dL). For values ≤100 mg/dL, the slope was 0.94 (R²= 0.92). No difference was found betweenthe 2 strip lots (P>0.05; Wilcoxon test).

CONCLUSIONS

StatStrip is suitable for measuring GLU in CSF. The use of this widely deployed instrument in our hospital departmentsfor this new application will provide critical, clinically actionable information in a timely manner. This will be especiallyuseful to pediatricians, neonatologists and in emergency departments, and offers the additional advantages of otherpoint-of-care tests, including small sample volume, ease of use and rapid results (6 seconds).


W150


A COMPARISON STUDY BETWEEN ROCHE ACCU-CHEK AND ROCHE COBAS 8000.

L. Galla 1, L. Sciacovelli 1, A. Padoan 1, M. Plebani 11Department of Laboratory Medicine, University-Hospital of Padova, Italy

BACKGROUND-AIM

The provision of health care has been profoundly changing in recent years, including through the Point Of Care Testing(POCT), to respond effectively to a patient-centred concept of care. Various esteemed organisations, both national andinternational, have defined guidelines for the correct management of POCT, but continuous monitoring of POCT dataremains essential to ensure the quality and reliability of the data obtained. This work aimed to verify the comparabilityof glucose values between those obtained in the analysis laboratory with Roche Cobas 8000 and those obtained withthe POCT device Roche Glucometer Accu-Chek® Inform II system.

METHODS

A total of 685 laboratory glucose samples obtained with Cobas 8000 modular analyzer and 708 glucose samplesobtained with Accu-Chek Glucometer between January to March 2019, 2020 and 2021 were collected. The agreementbetween glucose concentration (mmol/L) acquired on the same day and close of the same hour, with Roche Cobas 8000and with Accu-Chek has been evaluated with the Bland-Altman analysis and Passing-Bablok regression analysis. Inaddition, data were analysed based on their distribution into concentration ranges.

RESULTS

The Bland-Altman analysis highlights that the Bias% was not significant in all the three years taken into account(2019: 0.813%, 95%Cl: -0.504 to 2.130; 2020: 0.561%, 95%Cl: -1.899 to 3.022; 2021: 0.569%, 95%Cl: -0.611 to 1.749).However, a constant and proportional error was found through the years with the Passing-Bablok regression analysis,with respectively the following regression equations: 2019 (y=0.309 + 0.951 x), 2020 (y=0.622 + 0.899 x) and 2021(y=0.426 + 0.941 x). Analyzing the glucose results on the basis of concentration ranges, a consensus between results isfound, for example, for glucose concentration included between 6.99 mmol/L and 11.1 mmol/L: 46.56% in 2019, 37.7%in 2020 and 39.15% in 2021 of the laboratory results; 45.1% in 2019, 38.5% in 2020, and 39.5% in 2021, for POCT results.

CONCLUSIONS

The methods analyzed appear to be comparable, with no significant differences in Bias% through the years. However,the Passing-Bablok regression analysis reveals a constant and proportional error between laboratory results andPOCT results. These results underline the importance of continuous and dedicated monitoring of POCT devices, in allanalytical phases, to managing errors and the reliability of POCT results.


W151


HEMOLYSIS DETECTION WITH THE H-10 HEMCHECK DEVICE IN WHOLE BLOOD AND PLASMA STAT SAMPLES

A. Garcia-Osuna 1, A. Betbese 4, N. Benitez-Falga 1, M. Aceiton-Zabay 1, A. Mateos-Davila 4, M.I. Mirabete-Rodriguez4, H. Duhalde 2, J. Averstad 3, J. Ordonez-Llanos 11Department of Clinical Biochemistry (Hospital de la Santa Creu i Sant Pau, Barcelona, Spain)2Department of Health Sciences, Faculty of Health, Science, and Technology (Karlstad University, Karlstad, Sweden)3Hemcheck (Karlstad, Sweden)4Intensive Care Unit (Hospital de la Santa Creu i Sant Pau, Barcelona, Spain)

BACKGROUND-AIM

Hemolysis in blood samples causes analytical errors and can induce mistaken clinical decisions. It is only detectablevisually or photometrically by autoanalyzers after sample centrifugation. However, hemolysis is undetectable in wholeblood (WB) in the increasingly utilized point-of-care (POC) devices.A recently developed POC device (H-10, Hemcheck) allows for photometric detection of hemolysis (as free hemoglobin-fHb) in WB samples in syringes and tubes prior to their centrifugation. The fHb concentration is translated into abinary hemolytic index (HI).

METHODS

We aimed first, test the concordance between the HI result reported by the H-10 device in WB and an autoanalyzer(Alinity c) in plasma lithium-heparinized tubes collected from our ordinary clinical wards and second, use the H-10 forknowing the frequency and potential consequences of hemolysis in syringes obtained for blood gas analysis in ourIntensive Care Unit (ICU).Concordance between the H-10 and the Alinity was tested at the clinical lab by two technicians in 860 blood samplesin lithium heparin tubes; the frequency of undetected hemolysis in syringes was analyzed in 929 samples processed inPOC systems (ABL90 Flex) by the ICU personnel. We considered a fHb ≥1.0 g/L as a significant interference, and hencethe cut-off configured in and used by the H-10 device.

RESULTS

In plasma, a fHb ≥1.0 g/L was detected in 3.37% of samples by the Alinity and in 3.02% by the H-10, with 4 falsenegatives and 1 false positive in the latter. Thus, H-10 showed 86.2% sensitivity, 99.9% specificity and 96.2% positiveand 99.5% negative predictive values compared with the Alinity. As an average, HI result was obtained 44 minutesbefore by H-10 than by Alinity. In syringes, hemolysis was detected in 10.9% of venous and in 8.1% of arterial samples.Mean K+ was significantly higher (4.32 vs 4.06 mmol/L; p<0.001) and elevated pCO2 values (≥48 mmHg) were morefrequently observed in hemolyzed (47%) than in non-hemolyzed (36%) samples.

CONCLUSIONS

We conclude that the H-10 device had a similar performance as the autoanalyzer to detect hemolysis and detects ahigh prevalence of hemolysis in syringes. The rapid and easy detection of hemolysis by the H-10 system previous tosample centrifugation or analysis will avoid false results and erroneous clinical decisions and save wasting of sampleand reagents.


W152


PROENKEPHALIN (PENKID) AS A BIOMARKER FOR ACUTE KIDNEY INJURY AND ITS ANALYTICAL PERFORMANCEON APOINT OF CARE PLATFORM

D. Gruson 1, V. Carbone 1, K. Szczesna 2, J. Löffler 2, A. Schulze 2, D. Bergmann 2, T. Hartmann 21Cliniques Universitaires St-Luc2SphingoTec GmbH

BACKGROUND-AIM

Acute kidney injury (AKI) is defined as a rapid decline in renal filtration function. AKI is frequent in critically ill patients,and biomarkers can facilitate its early diagnosis and its prevention. Proenkephalin (penKid) is a stable opioid peptidecleaved from the same precursor of endogenous opioids as enkephalins. penKid has been identified as a functionalbiomarker for kidney function and its plasma concentration are inversely correlated to the glomerular filtration rate(GFR). The aim of our study was to evaluate the reliability of a new point-of-care testing (POCT) system for measuringpenkid with a short turnaround time of analysis (TAT).

METHODS

Inter-assay imprecision of the IB10 sphingotest® penKid® assay was assessed by repetitive measurement of humanEDTA plasma pools. Method comparison was performed with the reference to Immunoluminometric assay (ILMA) bymeasuring 50 plasma samples. Correlation and agreement between methods were tested using Bland-Altman plot andPassing Bablok linear regression. Usability of the POCT instrument was assessed by the multidisciplinary staff handlingthe POCT device through a dedicated questionnaire.

RESULTS

The inter-assay CVs of the IB10 sphingotest® penKid® test were 19.7% for a concentration of 113 pmol/L and 9.3%for a concentration of 307 pmol/L. The IB10 sphingotest® penKid® was significantly correlated to the ILMA (r=0.97).The Passing-Bablok linear regression showed a slope of 0.905 and an intercept of 7.49. The Bland-Altman plot showeda mean bias of 4.93% pmol/L between the two methods. The POCT usability questionnaire was overall satisfying. TheTAT analysis with the IB10 sphingotest® penKid® was below 20 minutes.

CONCLUSIONS

Our results showed a good analytical performance of the IB10 sphingotest® penKid® combined with a short TAT. Inaddition, an excellent correlation with the reference method was observed and the usability of the instrument wasevaluated as satisfactory by high majority of the users.


W153


NOVA BIOMEDICAL STAT STRIP XPRESS2 LACTATE METER VERIFICATION IN TARTU AMBULANCE FOUNDATION

A. Ivanov 3, A. Kõrgvee 2, I. Pastarus 11Anesthesiology and Intensive Care Clinic, Tartu University Hospital, Estonia2Tartu Ambulance Foundation, Estonia3United Laboratories, Tartu University Hospital, Estonia

BACKGROUND-AIM

Sepsis is a sequela of severe infection characterized by a systematic inflammatory response. Mortality rates from sepsisare high. Early antibiotic therapy can improve clinical outcomes and should be given within one hour of suspectedsepsis. Prehospital assessment can potentially impact and improve outcomes and time to initiation of treatments.Ambulance Foundations (AF) are focused on rapid diagnostic and treatment of patients with acute and urgent illnessesand injuries, including patients with sepsis. The objective of this study was to verify the professional StatStrip Xpress2lactate meter (Nova Biomedical, USA) with Tartu AF. We assessed the performance of this lactate meter by using aspreadsheet program for estimating the bias between two methods using patient samples. The verification of precisionwas performed by using control samples (Nova Biomedical, USA).

METHODS

The Rapidpoint 500e (Siemens, USA) was used as a comparative method for the study. According to EP09 (CLSI), 40whole blood heparinized samples used in comparison. Precision of the lactate meter was determined by measuring oftwo levels of quality control samples in 5 replicates over 5 days (EP15, CLSI).

RESULTS

The linear regression analysis of the method comparison study demonstrated a slope and intercept of 1,28 and -0,93,respectively. The results correlated well (R2= 0,987) and demonstrated that StatStrip Xpress2 and Rapidpoint 500ehad no significant bias (p-0,536). The range of lactate was from 0,72 to 17,46 mmol/L, mean lactate concentrationwas 3,64 mmol/L and the mean bias was -0,10 mmol/L (-8,9%). The intralaboratory imprecision of lactate across twolevels (0,6/6,4 mmol/L) was 21,6/10,6 %. The control results corresponded to a confidence level of 95% was 0,67 ±0,28mmol/L and 6,12 ± 1,29 mmol/L.

CONCLUSIONS

The StatStrip Xpress2 Lactate meter demonstrated a close correlation to the laboratory method. Precision of the meterexceeded the manufacturer’s claims, but was below biological intraindividual variation (CVi-27,3%). We conclude thatthe Nova StatStrip Xpress2 Lactate meter is suitable for use in the Tartu Ambulance Foundation. Prehospital lactatemeasurement improve outcomes by early identification of sepsis and initiation of treatment within one hour.


W154


ASSESSING THE ACTUAL ACCURACY OF BLOOD STATSTRIP (NOVA) GLUCOSE METERS.

J. Jacobs 1, J. Hotton 1, N. Miller 1, C. Lhoir 1, L. Brassinne 2, M. Cauchie 2, A. Mairesse 1, F. Vanneste 11Department of Biochemistry, Cliniques de l'Europe, Brussels, Belgium2Department of Microbiology, Cliniques de l'Europe, Brussels, Belgium

BACKGROUND-AIM

The StatStrip Glucometer (Nova Biomedical,USA) are cleared by US FDA and Health Canada for use with all patients,including critically ill. They have proven to be an accurate blood glucometer, but every laboratory needs to ensure thatthe glucometer used in the hospital is and stays fit for purpose. In hospitalized patients, decision are often based onmeasurements with these glucometers, and predefined and strict accuracy criteria are used. To follow the performanceof our glucometers we continuously extract data of our laboratory information system (LIS) to evaluate the actualaccuracy of our Statstrip glucose meters compared to our gold-standard, the Cobas (Roche) core lab analysers.

METHODS

Data from 1608 fasting patients having concurrent capillary and venous blood glucose measurements are compared.The capillary samples are analyzed immediately with the Statstrip glucometer (Nova biomedical, USA) by trainednurses. The venous samples are taken by the same nurses. The venous samples follow the routine pre-analyticalprocess to be afterwards analyzed on the core-lab Cobas (Roche) analyzers. The bias between Statstrip and Cobas iscalculated. Biaises were interpreted using the predefined accuracy Clinical and Laboratory Standards Institute (CLSI)POCT12-A3 criteria ( ( 1) 95% of samples should be within ±12.5 mg/dL for reference glucose values under 100.0 mg/dLand within ±12.5% for reference glucose values over 100.0 mg/dL and 2) 98% of samples should be within ± 15.0 mg/dLfor reference glucose values under 75.0 mg/dL or wihtin ± 20% for the reference glucose for values over 75.0 mg/dL)).

RESULTS

The mean bias calculated (-1.7 mg/dL for values under 100 mg/dL and -3.1% for values over 100mg/dL) met the CLSIPOCT12-A3 criteria but more than 5% of the samples did not meet the criteria ( [ -14.1 mg/dL- 10.7 mg/dL ] 95%Confidence interval for values under 100 mg/dL and [-24.9% - 18.7%] 95% confidence interval for values over 100mg/dL).

CONCLUSIONS

The Statstrip glucometer does not meet the POCT12-A3 criteria in our routine setting. As the mean bias lies wellwithin the predefined accuracy criteria but too many deviations are seen, causes, going from the preanalytical to thepostanalytical phase, needed to be identified to explain these sporadic inaccurate test results.


W155


AN INVESTIGATION TO ASSESS HOOK EFFECTS IN PREGNANCY TESTING KITS AND DEVICES

S. Jones 1, G. Davies 1, S. Jones 1, N. Blount 1, A. Thomas 11Weqas, Unit 6, Parc Ty Glas, Llanishen, Cardiff, CF14 5DU

BACKGROUND-AIM

IntroductionPregnancy testing kits or devices are used routinely in the UK both by health care professionals in clinical settings andby patients for home use. Research showed that the stated hook cut off varied greatly between kits / devices.AimsTo assess if there are any hook effects seen for Pregnancy Testing kits and devices (readers) registered on the WeqasUrine Pregnancy Testing EQA Programme (Proficiency testing).To assess if the observed hook effects match the manufacturers’ claims.To assess performance of Pregnancy Testing kits and devices at very high hCG concentrations.

METHODS

Urine was collected from healthy, non-pregnant female volunteers, filtered to 0.2µm and Gentamycin added tomaintain sterility. Intact hCG was added to the urine to a concentration of 700,000 IU/L. The pool was sent out to 180participants, selected to ensure similar numbers of results returned for each kit / device registered.Pools were also distributed at concentrations of 50 IU/L and 1000 IU/L, plus a negative sample.

RESULTS

Overall % positive rates for each pool were 1.3% for the negative pool, 78.6% for the pool at 50 IU/L, 98.7% for thepool at 1000 IU/L and 90.4% for the pool at 700,000 IU/L. Three kits had significantly lower % positive rates for the700,000 IU/L pool than the pool at 1000 IU/L.Of the 157 results returned for the pool at 700,000 IU/L:11 sites submitted a Negative result, across 6 kits / devices (7% of results).4 sites submitted a weak positive result across 3 kits (2.5% of results).142 sites submitted a positive result (90% of results).Only 1 reader submitted a negative result for the pool at 700,000 IU/L.

CONCLUSIONS

This study identified that hook effects were present in the kits / devices evaluated. Of the two kits and one devicethat showed significantly lower % positive rates for the pool at 700,000 IU/L, one kit insert did state a hook cut off of<500,000 IU/L. Data was not readily available from the other two manufacturers.The results demonstrate that it is imperative that users are aware of the limitations of the kits in use and that theycan easily identify hook cut off limits.


W156


INTERFERENCE TESTING ON THREE POINT-OF-CARE BLOOD GAS ANALYZERS

S. Lefever 1, T. Gajdos 1, N. Narinx 1, I. Van Hecke 1, A. Verdonck 11Point-of-Care testing, Laboratory Medicine, University Hospital Leuven

BACKGROUND-AIM

The purpose of this study was to evaluate several substances that potentially could influence co-oximetry andelectrolyte results on point-of-care blood gas analyzers. Interference by hemolysis, lipemia, benzalkonium chloride(BzCl) and methylene blue (MB) was investigated.

METHODS

Three point-of-care blood gas analyzers were evaluated in this study: (1) GEM® Premier™ 5000 (GP5000, Werfen,Barcelona, Spain), (2) RAPIDPoint® 500e (RP500e, Siemens Healthineers, Erlangen, Germany), and (3) ABL90 Flex Plus(Radiometer Medical ApS, Brønshøj, Denmark). Heparin anticoagulated whole-blood samples from healthy volunteerswere collected by venous puncture, after which they were spiked with the interfering substance of interest. Interferenceof hemolysis was studied by spiking a freeze-thawed whole-blood sample in fresh whole-blood samples to reachdifferent degrees of hemolysis. Lipid interference was investigated by spiking SmofLipid® (Fresenius Kabi, BadHomburg, Germany) in different concentrations (0.5, 1.0 and 2.0%) in whole blood. BzCl interference was tested in threedifferent concentrations (5, 10 and 20 mg/L) in whole blood. MB was spiked in four different concentrations (5, 10, 40and 80 mg/L) in whole blood. Samples were offered in triplicate to each blood gas analyzer for every concentration ofthe interfering substance in all spiking experiments, together with a blank whole-blood sample.

RESULTS

All analyzers experienced equal positive interference of hemolysis on potassium determination (up to +54%). BzClpositively interfered with the sodium measurement on ABL90 (up to +11%) and RP500e (no result), but not onthe GP5000. Small interferences due to lipemia were observed on co-oximetry results on all analyzers, while onlyhematocrite on GP5000 was positively influenced by lipemia (up to +17%). MB influenced co-oximetry results equallyon all three analyzers and sodium results on ABL90 (up to +13%) and RP500e (up to +9%).

CONCLUSIONS

All blood gas analyzers are influenced to a greater or lesser extent by the tested interfering substances. Each laboratoryshould know for their own blood gas analyzer which interfering substances are important and to which extent, andshould inform their clinicians about this.


W157


EVALUATION CAPILLARY VERSUS VENOUS BLOOD FOR HBA1C ON THE LAB 001 POCT INSTRUMENT

E. Lenters-Westra 11Clinical Chemistry, Isala Hospital, Zwolle, The Netherlands

BACKGROUND-AIM

The aim of this study was to evaluate capillary blood versus venous blood for HbA1c on a new capillary electrophoresispoint-of-care (POC) analyser (ARKRAY’s The Lab 001). The results were also compared with the IFCC Certified SecondaryReference Measurement Procedure (SRMP) Tosoh G8.

METHODS

From 62 donors with HbA1c values over the clinical range, capillary and venous blood was drawn. Capillary and venousblood were analysed on 2 The Lab 001 instruments and in duplicate on the Tosoh G8.

RESULTS

Deming regression analysis between capillary and venous blood showed a slope of 1.018 (1.004 to 1.032) and anintercept of -0.6689 (-1.428 to 0.09026). Bias at 30, 48 and 75 mmol/mol was 29.9 mmol/mol (28.48 to 30.25), 48.2mmol/mol (47.92 to 48.44) and 75.7 mmol/mol (75.23 to 76.10). Deming regression analysis between the Tosoh G8 andThe Lab 001 capillary blood showed a slope of 0.9479 (0.9279 to 0.9679), an intercept of 1.248 (0.2255 to 2.271) anda bias of 29.7 (29.20 to 30.17), 46.8 (46.44 to 47.05) and 72.3 (71.73 to 72.9). Deming regression analysis between theTosoh G8 and The Lab 001 venous blood showed a slope of 0.9647 (0.9388 to 0.9906), an intercept of 0.6017 (-0.7034to 1.907) and a bias of 29.5 (28.96 to 30.13), 46.9 ( 46.58 to 47.24) and 73.0 (72.20 to 73.71).

CONCLUSIONS

Statistically there was a difference between capillary blood and venous blood analysed on The Lab 001 as the slope of1.000 and bias at 75 mmol/mol were not within the 95% confidence intervals. However, these differences were so smallthat clinically seen there was no difference. Compared to the Tosoh G8 these differences were larger and statisticallysignificant, showing a negative bias particularly for very high HbA1c values (> 70 mmol/mol). The fast measurementtime, small sample volume and ease of use whilst giving clinically accurate results, makes The Lab 001 analyser a verysuitable platform to be placed in a diabetes facility or other appropriate point of care locations.


W158


POINT-OF-CARE TESTING IN THE EMERGENCY DEPARTMENT: IMPROVING PATIENTS' OUTCOMES

M. Jimenez-Barragan 1, C. Sanchez-Mora 1, P. Fernandez-Riejos 1, S. Martin-Perez 1, S. Fuentes-Cantero 2, A. Leon-Justel 11Clinical Biochemistry Department, Virgen Macarena University Hospital, Seville, Spain2Laboratory Medicine Department, Riotinto General Hospital, Huelva, Spain

BACKGROUND-AIM

Overcrowding in the Emergency Department (ED) is a global health problem. It’s necessary to find new strategiesto reduce length of stay (LOS) in this area and thus optimize workflows and use of resources. Point of Care Testing(POCT) can be a very useful tool to achieve these objectives, since, as has been observed in different studies, laboratoryturnaround time (LTAT) has a substantial influence on the total LOS in ED. In this study, we aim to demonstrate that theuse of POCT devices reduces total ED LOS in patients with intermediate priority and frequent reasons for consultation.

METHODS

We conducted a prospective cluster-randomized study to assess the impact of POCT in patients’ outcomes. The studywas carried out in February 2020 in the Virgen Macarena University Hospital (Seville, Spain). 208 Emergency SeverityIndex 3 (ESI3) patients (age older than 16) with the most common pathologies seen at this priority level in theED of our hospital (respiratory, genitourinary and abdominal) and with the need for analytical tests included in asimple analytical profile (haemogram, creatinine, urea, sodium, potassium, urianalysis, blood gas and/or InternationalNormalized Ratio) were studied in two different groups, interventional arm (laboratory analyses performed on POCTanalyzers implemented in ED; 101 patients) or control arm (central laboratory; 107 patients). Statistically, distributionswere examined using the Shapiro-Wilk test and as data had non-normal distributions, the Mann-Whitney test wasapplied to assess whether there were statistically significant differences (p < 0.05) in the ED LOS and LTAT betweenboth groups.

RESULTS

Results are expressed in median and interquartile range:Interventional armLOS: 155.00 minutes [114.00-223.00]LTAT: 22.65 minutes [16.82-30.27]Control armLOS: 239.00 minutes [198.00-300.00]LTAT: 94.18 minutes [72.27-121.33]pLOS: 0.001pLTAT: 0.001

CONCLUSIONS

We found a significant reduction in LTAT of 71.53 minutes and in LOS of 84.00 minutes. So, although further studiesare necessary, we think that implementation of POCT devices in the ED for ESI3 patients’ care can be a very useful toolto reduce LOS and, consequently, optimize resources, while improving workflows and management of all patients inthis area.


W159


COMPARATIVE STUDY OF HAEMOGLOBIN MEASUREMENT BETWEEN TWO POINT-OF-CARE TESTING DEVICES

M. Jimenez-Barragan 1, S. Martin-Perez 1, S. Fuentes-Cantero 2, C. Sanchez-Mora 1, P. Fernandez-Riejos 1, A. Leon-Justel 11Clinical Biochemistry Department, Virgen Macarena University Hospital, Seville, Spain2Laboratory Medicine Department, Riotinto General Hospital, Huelva, Spain

BACKGROUND-AIM

The use of Point-of-Care Testing (POCT) devices has become one of the most interesting tools in the EmergencyDepartment (ED). To ensure the analytical quality of the results of a POCT device that is intended to be implementedand its interchangeability with the reference analyzer, a method comparison study must be carried out between bothas part of the evaluation.In this study, we aim to verify the interchangeability of two analysers for the measurement of haemoglobin, which isa basic parameter handled in ED.

METHODS

A study was carried out to compare haemoglobin measurement methods in the Virgen Macarena University Hospital(Seville, Spain) between ABL90 FLEX PLUS (Radiometer; BrØnshØj, Denmark) and pocH-100i (Sysmex; Norderstedt,Germany) analysers, the latter being the reference analyser in the POCT setting of ED of our hospital.Haemoglobin was determined in parallel in a total of 110 whole blood samples from 55 patients attended in ED of ourhospital. Samples were collected in a venous blood gas syringe if processed in ABL90 FLEX PLUS or in a haemogramtube if processed in pocH-100i.The analysis was carried out with the statistical packages MedCalc and SPSS Statistics 25, comparing the results byPassing-Bablok regression and determining Spearman rank and intraclass correlation coefficients.

RESULTS

The results and their 95% confidence interval (95%CI) obtained were:Spearman rank correlation coefficient: 0.953 (0.921-0.973)Passing-BablokSlope: 1.0800 (0.9368-1.1286)Intercept: -0.5521 [(-0.6414)-0.6452]Intraclass correlation coefficient (R): 0.932 [(-0.061)-0.984]Spearman rank correlation coefficient was ~1 (p<0.01) and there were no significant deviations from linearity (pCusum<0.05).Passing-Bablok regression showed no systematic differences of constant type between the two analysers (95%CI ofthe intercept contains the value 0) or of proportional type (95%CI of the slope contains the value 1).R was ~1, but with low precision (95%CI).

CONCLUSIONS

A strong and positive correlation is observed according to Spearman rank correlation coefficient. Passing-Bablokregression indicates that methods are comparable for haemoglobin measurement and according to the intraclasscorrelation coefficient, the agreement is very good (R> 0.90), but there is very low precision.


W160


EVALUATION OF NEW POINT OF CARE TESTS PANEL FOR THE ASSESSMENT OF LIVER FUNCTION

C. Minea 1, V. Carbone 2, D. Gruson 21Department of Clinical Biochemistry, Hospital Universitario Fundación Jiménez Díaz, Madrid, Spain2Department of Clinical Biochemistry,Cliniques Universitaires Saint-Luc, UCLouvain (CUSL), Avenue Hippocrate 10, 1200,Brussels, Belgium

BACKGROUND-AIM

The use of point-of-care testing (POCT) is increasing because of its capacity of offering accurate and rapid diagnosis.Clinical laboratories and specialists in laboratory medicine play an important role for the implementation andvalidation of POCT instruments. The aim of our study was to determine the analytical performances of a recentlydeveloped POCT device for the measurement of tests related to liver function.

METHODS

We evaluated the performances of liver panel (Albumin, Alkaline phosphatase, AST, ALT, GGT, Direct Bilirubin, Totalbilirubin, Indirect Bilirubin, Glucose, Total Protein and Globulin) performed on the LINX EVO® POCT device (MenariniDiagnostics). The imprecision was determined with the Bio-Rad Liquichek Unassayed Chemistry Control. Methodcomparison was performed with Cobas® 8000 analyzer. Samples from twenty healthy volunteers were used to verifythe reference intervals. Furthermore, practicability assessed by the staff handling the POCT device through a dedicatedquestionnaire.

RESULTS

The imprecision observed for all the liver panel tests was matching the criteria provided by Westgard table. The onlyexception was the globulin with an observed imprecision of 6.3% and a criterion of 5.7%. Method comparison showedan excellent agreement with the routine laboratory method for all tests with however weaker agreement for total anddirect bilirubin. The verification of reference intervals showed that more than 90% of the healthy volunteers valueswere included into the reference interval claimed by the manufacturer with the exception of glucose and globulin whereadjustments will be needed. The POCT practicability questionnaire was overall satisfying for users and the rapidity ofmeasurement was confirmed.

CONCLUSIONS

Our study showed overall very good analytical performances for the liver panel performed on the LINX EVO® POCTinstrument. The user friendliness and rapidity of the system was also confirmed. To overcome the limited bias withroutine laboratory assays, the validation of local reference intervals and decision limits remain mandatory.


W161


ASSESSMENT OF NOVEL PANEL OF POINT OF CARE TESTS TO EVALUATE RENAL FUNCTION

C. Minea 1, V. Carbone 2, D. Gruson 21Department of Clinical Biochemistry, Hospital Universitario Fundación Jiménez Díaz, Madrid, Spain2Department of Clinical Biochemistry,Cliniques Universitaires Saint-Luc, UCLouvain (CUSL), Avenue Hippocrate 10, 1200,Brussels, Belgium

BACKGROUND-AIM

The use of point-of-care testing (POCT) is increasing because of its capacity of offering accurate and rapid diagnosis.Beside all the potential advantages that POCT devices offer, it is important to validate the reliability and quality of thesystems. The aim of this study was therefore to evaluate the analytical performances of a recently developed POCTdevice for the measurement of tests related to renal function.

METHODS

We evaluated the performances of renal panel (albumin, creatin kinase, creatinine, calcium, glucose, chlorine,phosphorus, potassium, sodium and GFR) of the LINX EVO® (Menarini Diagnostics) POCT instrument. The imprecisionwas determined with quality control materials. Method comparison was performed with Cobas® 8000 analyzer.Samples from twenty healthy volunteers were used to verify the reference intervals. Practicability of system was alsoreview by the staff handling the POCT device through a dedicated questionnaire.

RESULTS

The imprecision was estimated for each parameter and compared to the references provided by Westgard table.Albumin, creatin kinase, glucose and phosphorus were falling into the criteria whereas creatinin, calcium, chloride,potassium and sodium were slightly above. Method comparison showed a good agreement with the routine laboratorymethod with however weaker agreement for creatinin, calcium, chloride and potassium. The verification of referenceintervals showed that more than 90% of the healthy volunteers value were included into the reference interval claimedby the manufacturer with the exception of glucose, calcium, phosphorus and GFR where adjustments will be needed.The POCT practicability questionnaire was overall satisfying for users and the rapidity of measurement was confirmed.

CONCLUSIONS

Our study showed overall good analytical performances for the renal panel performed on the Linx Evo POCT instrument.The user friendliness and rapidity of the system was also confirmed. To overcome the limited bias with routinelaboratory assays, the validation of local reference intervals and decision limits remain mandatory.


W162


THE TRAINING STRATEGY IN POCT GOVERNANCE

D. Nigris 1, F. Cimolino 1, A. Duic 11University Hospital Udine - Clinical Laboratory

BACKGROUND-AIM

The diffusion of SARS-CoV-2 pointed out several problems in the Italian health system, highlighting the need toimprove new fields, like for example: the implementation of the POCT’s governance in the public and private Hospitaldepartments.Nowadays, the development of rapid tests is constantly increasing, but on the other hand there is an uncontrolledcommercialization of unsafe products. So, the diffusion of these tests, without an appropriate management, could beeconomical expensive but also really dangerous in the Hospital departments.

METHODS

It’s really necessary to develop an informatic network in order to improve the governance of the POCT. In fact, in thisstudy we decided to create several levels of trainings, in order to answer to the different problems in our Hospital andUniversity Health System which includes a wide area from the mountains to the sea.

RESULTS

In details we developed:1. POCT lab technician Specialist Training: this pool of lab techician will manage in the future the administration andmanagement of the POCT’S instruments, they will be concerned with the execution of the Internal/External QCs;2. Lab technician Training: in this training the previous Specialists will introduce to the lab technicians the innovationof the POCT’S government and instrument, just to let them know the advantages and application of POCT in the criticalareas;3. Users training: this training it’s addressed to the other health professional (e.g nurses) in Hospital to let them knowhow to use the POCT instruments, but also the right way to execute the different types of sample in the medical fields(e.g. pathology, microbiology).4. Creation of a University First level Master: in this Master the aim is to improve knowledges and skills of the POCTapplied to clinical pathology, microbiology/virology through a new cross-curricular formative course addressed tohealth professional, such as pharmacist, clinical engineer and obviously lab technicians.

CONCLUSIONS

To conclude, this study highlighted the need to introduce a new training approach, by using the new technologicalplatforms (e.g videoconferences, apps), especially after the diffusion of SARS-CoV-2, in order to reduce the face-to-face contacts.


W163


IMPACT OF THE COVID-19 CRISIS ON THE POINT-OF-CARE BLOOD GASES MANAGEMENT (GEM PREMIER SERIES,WERFEN) IN AN ISO 22870 ACCREDITED LABORATORY IN PARIS (FRANCE)

P. Pernet 1, B. Beneteau-Burnat 1, E. Maury 2, P. Vitry 2, L. Satre-Buisson 3, F. Pinto 3, P. Leveque 1, C. Aussel 1, M.

Vaubourdolle 11Biochemistry Laboratory, Department of Biochemistry-Hormonology, Saint-Antoine Hospital, DMU BioGeM, AP-HPSorbonne University, Paris, France2Medical Intensive Care Unit, Saint-Antoine Hospital, DMU DEMIURGE, AP-HP Sorbonne University, Paris, France3Surgical Intensive Care Unit, Saint-Antoine Hospital, DMU DREAM, AP-HP Sorbonne University, Paris, France

BACKGROUND-AIM

Our laboratory is accredited for point-of-care (POC) blood gases activities according to the ISO 22870 standard. When,in March 2020, the Covid-19 crisis hit France, risk assessment was done to adapt POC management to the setting upof 3 new dedicated Covid-19 intensive care units.

METHODS

We used our change management procedure based on risk assessment management (CLSI EP-23) and prioritization ofrisks (criticality scale) to reveal new risks to take into account:Material: - Insufficient number of blood gases devices. - Shortage of reagents requiring anticipating the potential peakof analysis (units full of ventilated patients).Manpower: - Newly trained-empowered people recruited for the Covid-19 crisis requiring to derogate from the usualtraining procedure to increase the rate of operational users. - Staffing shortage in the clinical units or in the laboratorydue to illness.Methods: - Reduced initial analyzer performance check for quick commissioning. - Necessity of indicators to monitorthe impact of derogations to usual procedures and to verify the adequacy with the clinicians needs.Environment: - Potential impact of SARS-CoV-2 on analyzers management (device contamination, waste, protectionof users).

RESULTS

Material: One GEM Premier 4000 was put back into service and 2 additional GEM Premier 5000 (Werfen) were orderedand put into service the day of reception 9 days later. Reagent orders were tripled to avoid shortage due to potentialfuture manufacturer’s deficiency. Eventually, POC blood gases activity increased +300% at the peak of the crisis in Aprilwithout any particular problem.Manpower: 35 new users were trained-empowered with a quick-training procedure. A punctual lack of trained usernever happened and daily monitoring of rejected analysis (new indicator) showed even better results than expected.Methods: Indicators allowed to verify that specific requirements of ISO 22870 were still achieved.Environment: No special procedure was required for the analyzer itself aside the general procedure for COVID-19clinical unit management.

CONCLUSIONS

Our change management procedure allowed our ISO 22870 accredited laboratory to add these new locations/POCanalyzers to our scope of accreditation during the peak of the COVID-19 crisis.


W164


LONG-TERM PERFORMANCE OF POINT-OF-CARE HEMOGLOBIN A1C ASSAYS

W. Masri 1, E. Plouvier 1, M. Bendaoud 1, Y. Costa 11biochemistry department, Grand Hôpital de l'Est Francilien - site de Meaux (EST-Ile de France) – France

BACKGROUND-AIM

With evolving laboratory technology and, in particular, miniaturization of processes and reduction in required samplesize, point-of-care (POC) measurement has become a key modality to provide rapid information needed for clinicaldecisions. POC testing of HbA1c is used as a time-efficient tool to improve treatment and management planning fordiabetes in the pediatric clinical service. We assessed the reliability of POC HbA1c assay techniques in a real world clinicsetting over a 4-year period, comparing the results to 2 central laboratory techniques.

METHODS

We compared the DCA Vantage® from Siemens™ immunoassay-based POC techniques to the G8® from Tosoh™ andVariant II® from Bio-Rad™ ion-exchange high-performance liquid chromatography (HPLC) central laboratory methodsin a study lasting 4 years. In total, 80 patients in pediatric service participated in this study diagnosed with type 1diabetes and undergoing treatment. A venous blood sample was taken for routine HbA1c analysis by the diagnosticlaboratory method, a blood sample was taken from the fingertip For the POCT HbA1c at the same time. The HbA1cresults of three methods were compared and analyzed with a Bland-Altman agreement plot.

RESULTS

The DCA Vantage® also has good concordance HbA1c ranged from 5 –14�% with the central laboratory G8® (Bland–Altman plots Coef R2: 0.997) and Variant II® (Bland–Altman plots Coef R2: 0.995). Despite high correlations amongthe 3 techniques, there were variable differences obtained over time. The Bio-Rad and Tosoh values varied from 0.1 to0.2% higher than the DCA values but the difference is clinically insignificant.

CONCLUSIONS

POCT HbA1c measurements using the DCA Vantage system, compared to the standard laboratory method, are highlysensitive and accurate, as well as being less time-consuming and costly, with a more convenient blood collection.Therefore, this POCT system is a suitable tool for diabetes screening and management in the service pediatric. However,the linearity limit for the DCA Vantage® is 14% above this value, the HbA1c must be carried out in the laboratory.


W165


ACCURACY OF POINT-OF-CARE OF MEASURING NEONATAL BILIRUBIN CONCENTRATION

W. Masri 1, H. Broutier 1, H.A. Pichon - Cung 1, E. Plouvier 11biochemistry department, Grand Hôpital de l'Est Francilien - site de Meaux (EST-Ile de France) – France

BACKGROUND-AIM

The Point-of-Care (POCT) ABL 90 Flex plus® blood gas analyzer has the advantage of measuring total bilirubin (tBil)with a capillary sample and a very small volume of blood in neonatology. Since the marketing of this machine, noorganization has offered an external quality control (EQC) to estimate the accuracy of the tBil. The objective of ourstudy is to evaluate the accuracy of total bilirubin measured by the POCT ABL 90 Flex plus® from Radiometer™ usingEQC from Probioqual for Cobas 6000®.

METHODS

45 EQC for tBil between 2018 and 2021 intended for the diazo method in cobas 6000 c 501® from Roche™ wereanalyzed by co-oximetry in POCT ABL 90 Flex plus® from Radiometer™ in the neonatology department. The sampleswere measured on the same day on two automates. The results were processed via the guidelines method validationaccording to ISO 22870 and were compared and analyzed with a Bland-Altman agreement plot.

RESULTS

The tBil concentrations covered a range of 16 to 322 µmol/L, a good correlation between the two techniques with acorrelation coefficient of 0.993 and the regression equation y: 1.0163X- 1.2738.For all tBil concentration levels the observed deviation was in acceptable limits for the French Society of Clinical Biology.A small bias is observed for low values compared to the cobas peer group but it’s statistically and clinically insignificant.

CONCLUSIONS

In the absence of a tBil EQC on ABL90 Flex plus® from Radiometer™, this alternative measurement can be adapted bythe laboratory to assure the monitoring of the performance of this sensitive and important setting for the diagnosisof jaundice in newborns.


W166


ANALYTICAL VALIDATION OF PROCISE DX POC ANALYSER FOR THERAPEUTIC DRUG MONITORING OF INFLIXIMAB

I. Šahinović 1, V. Oršić-Frič 3, V. Konjik 4, J. Pavela 1, V. Borzan 2, V. Šerić 11Clinical Department for Laboratory Diagnostics, University Hospital Osijek, Croatia; Faculty of Medicine, J.J. StrossmayerUniversity of Osijek, Croatia2Department of Gastroenterology, Clinic of Internal Medicine, University Hospital Osijek, Croatia3Department of Gastroenterology, Clinic of Internal Medicine, University Hospital Osijek, Croatia; Faculty of Medicine, J.J.Strossmayer University of Osijek, Croatia4Department of Pediatric Gastroenterology, Clinic of Pediatrics, University Hospital Osijek, Croatia

BACKGROUND-AIM

Within the past 20 years the use of biologic drug infliximab (IFX) in inflammatory bowel disease (IBD) treatment hasbeen an important step forward in improving disease course and keeping the inflammation at remission for a prolongedperiod of time. Therapeutic drug monitoring (TDM) of IFX using the point-of-care (POC) methods showed to be able toproduce results fast enough to allow IFX dose adjustments prior to drug infusion. This study aimed to compare twoPOC tests, Procise IFX (Procise Dx) and Quantum Blue infliximab assay (Bühlmann Laboratories AG) used in UniversityHospital Osijek.

METHODS

Serum samples from 22 IBD patients on IFX maintenance therapy were collected immediately before drug infusion.Capillary blood was also collected from 6 IBD patients by finger prick using whole blood pipettes (Procise Dx).Analytical imprecision was assessed using the human serum sample obtained from a patient on IFX therapy with IFXconcentration 5,4 mg/L. For method comparison a Passing Bablok regression was used and for qualitative comparisonweighted kappa statistics was obtained after stratification of results by therapeutic range (<3 mg/l, ≥3 to 7 mg/L, and≥7 mg/L).

RESULTS

Within- and between-laboratory precision of Procise IFX assay was 3,9% and 1,4%, respectively. Passing Bablokregression analysis of IFX concentration in serum sample has showed proportional deviation between POC methods(y=0.416 (-0.654 to 1.279) + 0.753 (0.648 to 0.923)x). Good comparison has been observed for capillary blood measuredusing Procise IFX test and serum samples measured using Quantum Blue Infliximab assay (y=-2.732 (-4.315 to 1,867)+ 1.263 (0.278 to 1.566)x), although IFX concentration was lower in capillary blood. Classification of results accordingto therapeutic interval showed good agreement for serum samples as well as for serum and whole capillary bloodmeasurements (κ=0.79 (0.59-0.999) and κ=0.80 (0.45-1.00), respectively).

CONCLUSIONS

Procise IFX assay showed good analytical performances. Because of proportional differences, two POC tests can’t beused interchangeably for longitudinal TDM of infliximab in IBD patients. The use of whole capillary blood for TDM ofIFX should be evaluated in a larger number of subjects.


W167


DETECTION OF SEVERE IN VIVO HEMOLYSIS DURING EXTRACORPOREAL MEMBRANE OXYGENATION SUPPORTIN COVID-19 PATIENTS. RELATION OF PLASMA FREE HEMOGLOBIN WITH ENDOGENOUS CARBON MONOXIDEPRODUCTION.

M. Serres Gómez 1, M. Simón Velasco 1, K.L. Nanwani Nanwani 2, B. Fernández Puntero 1, M.J. Alcaide Martín 1, M.

Rodríguez Gutiérrez 1, M.P. Sanz De Pedro 1, A. García Muñoz 2, C. Gutiérrez Melón 2, M. Quintana Díaz 2, P. Millán Estañ2, R. Gómez Rioja 11Department of Laboratory Medicine, La Paz University Hospital, Madrid2Intensive Care Unit, La Paz University Hospital, Madrid

BACKGROUND-AIM

In the COVID-19 pandemic, the Extracorporeal Membrane Oxygenation (ECMO) treatment has proved its usefulness inpatients with acute lung injury. In vivo hemolysis (ivH) is one of its complications, characterised by peaks of plasmafree hemoglobin (fHb). Moreover, an increase in carboxyhemoglobin (COHb) has been observed in several cases due toHb metabolism by heme-oxygenase that releases carbon monoxide. The aim of this study is to evaluate the incidenceof ivH events and their relation to COHb in COVID-19 patients with ECMO treatment.

METHODS

From March 2020 to June 2021, 32 COVID-19 patients received ECMO in our Intensive Care Unit, with an averageduration of 4 weeks. Daily analytical monitoring was carried out including arterial blood gas test with cooximetry(ABL90 FLEX, Radiometer) and biochemical parameters (Atellica Solution, Siemens Healthineers), incorporating theestimation of fHb using quantitative hemolysis index (HI) (spectrophotometric measurement). Significant ivH wasconsidered with fHb>100 mg/dL, whereas severe ivH required fHb>500 mg/dL.Daily maximum value of HI and COHb was selected to analyse the results, obtaining a total of 825 pairs of results.These were grouped by weeks and the weekly mean was calculated for both to evaluate their correlation.

RESULTS

One fourth of the patients presented at least one peak of significant ivH, out of which 62,5% died. Severe ivH wasdetected in 4 patients (12,5%), associating a 50% mortality rate. Half of the hemolytic events required ECMO membraneor circuit exchange.The COHb mean value from all the results was significantly higher when compared to the upper reference limit of thenormal range in non-smokers (3,1% vs 1,5%, p<0,01). COHb>4% was detected in all the cases of severe ivH. The highestpeak of fHb was 1237mg/dL with a COHb of 7,1%.A positive correlation was observed between the weekly means of HI and COHb (Pearson coefficient r=0,47; p<0,01).

CONCLUSIONS

During ECMO treatment, episodes of ivH are very frequent, in which the increase of fHb is positively correlated withCOHb. Moreover, maintained high COHb levels in these patients imply higher toxicity and oxygen transport impairment.Thus, detection of ivH events using HI and Point-of-Care COHb can aid in better monitoring and management of ECMO-related hemolysis.


W168


METHOD COMPARISON OF SIEMENS HEALTHINEERS’ BLOOD GAS SYSTEMS FOR COMMON ANALYTES: EPOC BLOODANALYSIS SYSTEM VERSUS RAPIDPOINT 500E BLOOD GAS SYSTEM

S. Shum 1, T. Umscheid 11Siemens Healthineers

BACKGROUND-AIM

The objective of the study was to demonstrate harmonization between the Siemens Healthineers epoc® Blood

Analysis System for patient-side testing and the RAPIDPoint® 500e Blood Gas System designed for the Point-of-Care. All common analytes were evaluated to demonstrate correlation between the two systems and to minimize theopportunity for misinterpretation of test results that could adversely impact patient outcomes.

METHODS

A method comparison study was performed on the Siemens epoc Blood Analysis System (y) versus the RAPIDPoint500e Blood Gas System (x) following the study design and analysis in the CLSI EP09c guideline. The analytes assessedincluded blood gas (pH, pCO2, pO2), electrolytes (sodium, potassium, chloride, ionized calcium) and metabolites(glucose, lactate). A minimum of 100 heparinized whole blood samples per analyte were assayed on two of each system.

Correlation statistics including slope, intercept, coefficient of determination (r2), and biases at the medical decisionlevels were generated.

RESULTS

Correlation statistics and regression graphs will be presented for each analyte.

CONCLUSIONS

The method comparison demonstrated harmonization between the epoc Blood Analysis System and the RAPIDPoint500e Blood Gas System at the clinically relevant medical decision levels.


W169


USING BULL'S ALGORITHM FOR POCT ANALYSERS IN HAEMATOLOGY PATIENTS

M. Simac 1, L. Bilic-Zulle 11Clinical Department of Laboratory Diagnostics, Clinical Hospital Center Rijeka, Rijeka, Croatia

BACKGROUND-AIM

Bull's algorithm (XbarB protocol) is a methods used for calculating moving average as internal quality control inlaboratory. It uses batches of 20 continuous patient results. Truncation, and action and target limits must be set up.Any shift or trend of data out of the set target limit by more than the set action limit should be reviewed.POCT analysers are used by non-laboratory personnel at out-of-laboratory sites. More quality control procedures areneeded to detect system faults since supervision is not always possible by the overseer.

METHODS

Although some haemotology analysers have automatic moving average calculation within their software, the onewe use (pocH-100i (Sysmex, Germany)) does not, as such we used XbarB protocol to calculate it ourselves. Theparameters used to calculate moving average are erythrocyte constants, which, if there is no significant change inpatient population, varies up to 0.5% on a day-to-day basis. We set truncation limits for MCHC between 250 and 370g/L based on previous experience for sample acceptance; if lower or higher, the sample is rejected due to preanalyticalerrors (such as sample dilution or blood clots). The action limit was set to 3% to detect a system error of over 3%, andstable moving average as the target limit.

RESULTS

During January 2020 in Clinical Hospital Center Rijeka all patient results were collected from the Haematology DayClinic and processed Bull's algorithm to detect analyser system fault. There were 349 samples with 3 excluded due totruncation limits. One instance of moving average was detected exceeding the set action limit of 3%, on the day therewas a change of reagents lot.

CONCLUSIONS

To use Bull's algorithm, a population of patients is needed where similar results are expected for parameters which donot statistically differ day-to-day. This analyser and site were convenient due to the homogeneous patient population.XbarB protocol is useful for detecting system faults as shown by detecting a shift the day the reagents lot was changed,which is one of the expected reasons for a shift.


W170


EVALUATION OF THE RELIABILITY OF THE HEMATOLOGICAL DATA OF AN INSTRUMENT IN POCT

f. Spolaore 1, m. Fontanari 1, s. Dall'Agnol 1, a. Zelletta 1, l. Menghini 1, a. Anesi 11U.O. Multizionale Patologia clinica - Ospedale S Chiara Apss Trento

BACKGROUND-AIM

The widespread use of Point of Care Testing has allowed us to obtain many clinical / therapeutic needs in an emergencyregime, to provide accurate and rapidly available results. Decentralized diagnostics uses PoCT technology, currentlyrepresents one of the most suitable organizational solutions to support those situations in which it is necessary tohave a rapid laboratory diagnostic result, but validated and correlated and correlated with the data provided by theCentral Laboratory (Hub) , in compliance with the required quality standards. This need is evident above all in theEmergency / Urgency departments, such as the First Aid, intensive therapy and / or oncological departments whereclinical evaluations on patients must be rapid and decisive, according to a precise classification of the patient and anadequate therapeutic course.

METHODS

In our study we preliminarily assessed the reliability of the main haematological parameters and of the leukocyteformula of the PoCT ICON5 (A.DeMori) by analyzing 52 patients (26 F 26 M) chosen from the samples of thehaematological routine of the S Chiara (TN) laboratory - Clinical Multizonal Pathology Unit, aged 1-86 years, whocome from different parts and with different pathologies. To evaluate the feasibility of the project, we synchronouslyevaluated the new generation ICON 5 automatic hematology analyzer, capable of determining the haematologicalparameters with differentiation of the leukocyte formula in 5 populations and the instrument used in routine XN10(Dasit-Sysmex) user in our chain routine. The statistical analysis was carried out using the Bland Altman and PassingBabblok regression tests (Analyze.it).

RESULTS

From the still preliminary tests carried out to date, it is possible to deduce a good search between the two instruments,with a significant percentage Bias only for the parameter Hb. By way of example, we report four graphs of the statisticalprocessing of some parameters of the blood count and of the leukocyte formula.

CONCLUSIONS

We therefore hope that this tool can be of support for the emergencies above. A quick and reliable tool to support theneed for a quick and precise clinical classification, probably to be completed also with the evaluation by the laboratory


W171


MULTICENTRE EVALUATION OF A NEW STRIP-BASED BLOOD GLUCOSE SYSTEM (COBAS® PULSE, ROCHE DIAGNOSTICS)FOR NEAR-PATIENT TESTING IN CRITICAL AND NON-CRITICAL CARE SETTINGS

M. Goodman 7, N. Breitenbeck 12, W. Clarke 5, C. Canada-Vilalta 15, D.D. Canepa 14, E. Ganser 13, B. Goldsmith 4, G. Headden10, D. Hoppensteadt 11, B. Klapperich 14, R. Matika 1, G. Miles 15, E. Moser 15, J.H. Nichols 3, Z. Rafique 9, A. Stankiewicz15, J. Swanson 6, N. Tran 2, A. Wu 16, B. Karon 81Department of Anesthesiology, University of Arizona Health Network, Tucson, AZ2Department of Pathology and Laboratory Medicine, University of California Davis Medical Center, Sacramento, CA3Department of Pathology Microbiology, and Immunology, Vanderbilt University Medical Center, Nashville, TN4Department of Pathology, Anatomy, and Cell Biology, Thomas Jefferson University, Philadelphia, PA5Department of Pathology, Johns Hopkins University School of Medicine, Baltimore, MD6Department of Pediatrics, University of Virginia Children’s Hospital, Charlottesville, VA7Department of Surgery, University of Cincinnati, Cincinnati, OH8Division of Clinical Core Laboratory Services, Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN9Emergency Medicine, Baylor College of Medicine – Ben Taub Hospital, Houston, TX10Emergency Medicine, Medical University of South Carolina, Charleston, SC11Molecular Pharmacology and Neuroscience, Loyola University Chicago, Chicago, IL12NB Research, Inc., Indianapolis, IN13Roche Diabetes Care Inc., Indianapolis, IN14Roche Diagnostics International AG, Rotkreuz, Switzerland15Roche Diagnostics Operations, Indianapolis, IN16University of California San Francisco General Hospital, San Francisco, CA

BACKGROUND-AIM

cobas® pulse is a new near-patient blood glucose testing system for use in professional healthcare environments. Thismulticentre study aimed to evaluate the performance of this system using arterial (A), venous (V), capillary (Cap) andneonatal heel stick (HS) whole blood samples.

METHODS

The performance of the cobas® pulse system was compared with that of a plasma-based hexokinase reference method

on the Roche cobas® 6000 system (GLUC3) and to that of Nova StatStrip® (Nova Biomedical). Accuracy was assessedby CLSI POCT12-A3 criteria: ≥95% of results within ±12 mg/dL (±0.67 mmol/L) of the reference method at <100 mg/dL(5.55 mmol/L) and ±12.5% of the reference method at ≥100 mg/dL (12/12.5 criteria); and ≤2% of results exceeding ±15mg/dL (±0.83 mmol/L) at <75 mg/dL (4.2 mmol/L) and ±20.0% of the reference method at ≥75 mg/dL (15/20 criteria).

RESULTS

Overall, 1142 samples (n=316 A, n=706 V, n=120 HS) were collected from 975 adult, paediatric and neonatal patientsin critical and non-critical care settings at 12 US sites. Cap measurements (n=702) were performed on samples from

117 non-critically ill adult subjects at one US site in a separate study. For the cobas® pulse system, 98.7%, 97.9% and95.0% of A, V and HS results, respectively, met CLSI 12/12.5 criteria at the 100 mg/dL breakpoint and 0.6%, 1.1% and0% exceeded the CLSI 15/20 cut-off at the 75 mg/dL breakpoint (vs the reference method), fulfilling the criteria. By

contrast, for StatStrip®, 87.7%, 92.5% and 92.5% of A, V and HS results, respectively, met CLSI 12/12.5 cut-off at the 100mg/dL breakpoint, which is below the required 95% threshold, and 0.9%, 0.9% and 1.7% exceeded the CLSI 15/20 cut-

off at the 75 mg/dL breakpoint, fulfilling the second criteria. For Cap results, the cobas® pulse system met CLSI criteriaat both breakpoints (99.4% were within 12/12.5 criteria at 100 mg/dL; 0.3% exceeded 15/20 criteria at 75 mg/dL),

whereas StatStrip® met CLSI 15/20 criteria (0.9% exceeded limits), but not 12/12.5 criteria (94.0% were within limits).

CONCLUSIONS

The cobas® pulse system met CLSI POCT12-A3 accuracy criteria for all A, V, Cap and HS samples, and the performance

exceeded that of StatStrip®.


W172


POINT OF CARE TESTING (POCT) IN DIABETES MANAGEMENT: RURAL AND REMOTE PRIMARY CARE SETTINGS

G.M. Varo Sánchez 1, M. Rico Rodríguez 1, M.J. Ruíz Márquez 1, I. Zakariya-Yousef Breval 11Análisis Clínicos, Hospital de Riotinto, Huelva, Spain

BACKGROUND-AIM

In recent years there has been an exponential development of POCT tests, the existing evidence shows that one of thekey advantages is the availability of these devices at the “bedside”, a fact that supports their use in rural and largehealthcare areas with population dispersion. Using POCT technology response times are shortened, which speeds upclinical decision-making and optimizes the therapeutic approach.The most recent recommendations of the POCT International Clinical Practice Guidelines, which mention the figure ofthe POCT Coordinator as key in the management and approach to the implementation of any biomarker, in a hospitalor out-of-hospital area, in order to create close collaborations between the Laboratory and Physicians, providing addedvalue to the POCT laboratory networks.The aim of this study was to evaluate the impact of the results between POCT stations, located in rural and remotedepartments, and Central Laboratory according to healthcare needs in diabetes management.

METHODS

This is a retrospective observational study, performed using information stored in the Laboratory Information System(LIS) with regard to critical result notifications reported between 01.01.2021 and 01.08.2021. The devices used wereprofessional glucometers Accu-chek Inform II by Roche. Laboratory technicians remotely verify and monitor all POCTsystems with the programme Cobas IT. Two levels of quality control for glucose are performed once a day and if it is notmade the department staff cannot perform any glucose test before quality control performance. Central Laboratorysupervises the quality control performance and gives a technical assistance when it is required. A list of critical valuesfor all POCT devices was prepared by the Central Laboratory in consultation with the attending Physicians. All POCTresults are released via the LIS to the patient sheet and become part of their clinical history.

RESULTS

It available 28 glucometers in 4 departments of Primary Care and there are 20 glucometers all over the Hospitalconnected by Wi-Fi. A total of 14129 samples and 4257 quality controls were processed, all of them accepted fromCentral Laboratory according to Westgard rules. Furthermore, 20 patients showed results of glucose lower than 40 mg/dl and 9 patients presented results of glucose upper than 600 mg/dl. Physicians activated protocol of critical values ofglucose in 37 patients. Additionally, during the period of this study, a total of 15 patients were diagnosed of Diabetestype 2 from Primary Care with POCT technology. Only 2 devices were disconnected due to technical assistance, but thenew devices were available 24 hours later by technical service.

CONCLUSIONS

In conclusion, the scope of this organizational model of POCT laboratory network is the improvement of qualityof patient care. This technology assists laboratory and non-laboratory staff who are working with POCT improvingthe management and safety patient of Diabetes. The integration between decentralized diagnostics and CentralLaboratory aims at guaranteeing that analytical quality is maintained, in order to ensure results reliability andprecision, and patient safety.


W173


USEFULNESS OF CARBOXYHEMOGLOBINEMIA VALUE IN THE DEVELOPMENT OF COVID-19 PNEUMONIA

G. Velasco De Cos 2, A. Maiztegi Azpitarte 2, L. Velasco Rodríguez 1, R.J. García Martínez 2, M. Iturralde Ros 2, J.M. Cerezo

Martín 3, N.H. Cahuana Santamaría 2, E. Pizarro Peña 1, B.A. Lavín Gómez 21Clinical analysis service, Hospital San Agustín, Linares, España2Clinical analysis service, Hospital Universitario Marqués de Valdecilla, Santander, Spain3Haematology service, Hospital Universitario Marqués de Valdecilla, Santander, Spain

BACKGROUND-AIM

Carboxyhemoglobin (COHb) is generated as a result of CO binding to the haemoglobin molecule, decreasing its oxygen-carrying capacity. Its value is increased in cases of intoxication, sepsis, haemolysis and severe inflammatory conditions.In the context of the current COVID-19 pandemic, the COHb value can be obtained relatively easily and inexpensivelyfrom patient blood gases. Our study aims to find out whether its elevation is proportional to the severity of the illness,assessing this according to the type of admission required. In addition we are going to evaluate if there is a correlationbetween the peak reached and the duration of the admission.

METHODS

An analysis was made of the COHb results obtained in our centre using the Radiometer ABL800 FLEX analyser in 140patients with SARS-Cov2 infection. The duration and type of admission were recorded from the clinical history: Non-admitted (NA), ICU and ward. In patients with multiple determinations, the highest result during the course of infectionwas selected. A Levene's test was performed (p=0.000), so the analysis was continued with the Kruskal-Wallis test. Weused the spearman’s Rho to test the relationship between the COHb and days of admission.

RESULTS

COHb values follow a normal distribution, finding significance with the Kruskal-Wallis test for non-admitted (x�=0.89± 0.35), ward-admitted (x�=1.47± 0.63) and ICU-admitted (x�= 1.92 ± 0.96). Pairwise comparison showed a p=0.007for NA-plant, p= 0.000 for NA-ICU and p= 0.023 for plant-ICU.Spearman's test showed a correlation between days of admission and COHb values (p= 0.000). With a ρ of 0.493.

CONCLUSIONS

COHb is a test available in most blood gas analysers and its determination by spectrophotometry does not usuallyimply an additional cost when performing the patient's blood gas analysis. Our group has found significant differencesin the values of this parameter depending on the severity of the infection. There is also a correlation with the durationof hospitalisation, although this does not show a good ρ. Based on these results, the COHb value provides an addedvalue to the determination of blood gases in COVID-19 patients that can help the clinician in assessing the severity ofthe condition. Further studies on the importance of COHb are needed to understand the role it plays in the developmentof infection.


W174


REPEATING THE CRITICAL VALUES OF COVID-19 PATIENTS ON POCT ANALYZERS EXPEDITES PHYSICIAN REPORTING

M. Žarak 1, A. Lukenda 1, M. Tomičević 1, S. Marević 1, L. Đerek 11Clinical Department for Laboratory Diagnostics, Dubrava University Hospital, Zagreb.

BACKGROUND-AIM

The monitoring of COVID-19 patients has become a global challenge in many clinical laboratories, and there is anincreasing incidence of critical values in laboratory results as compared to other patients in intensive care units. Criticalvalues represent those laboratory results that indicate a life-threatening condition of the patient, and it is commonpractice in many laboratories to repeat critical results before reporting the value to the physician. Such results shouldbe reported within 30 minutes of confirmation, and waiting for a re-measurement may delay reporting.

METHODS

Since it was noticed that COVID-19 patients had the highest percentage of critical values for potassium (K), sodium(Na), chloride (Cl) glucose (Glc), and lactate (Lac), the aim of this summary was to compare results obtained primarilyon a biochemical (Beckman Coulter DxC 700 AU) and repeatedly on a less time-consuming POCT (Siemens RapidLab1265) analyzer. For this purpose, 40 serum samples were used for the determination of K, Na, Cl and Glc and 10 plasmasamples for the determination of Lac. Statistical comparison of the results was performed using Passing-Bablock (forelectrolytes and Glc) and Deming regression analysis (for Lac).

RESULTS

Analysis of the results revealed neither proportional, nor constant error for K [y = -0.15 (-0.29-0.02) + 1.03 (0.99-1.07)x], Na [y = -4.08 (-15.47-5.19) + 1.03 (0.96-1.11) x], Cl [y = 5.00 (-2.96-5.00) + 1.00 (1.00-1.08) x], and Glc [y = -0.10(-0.24-0.04) + 1.00 (0.98-1.02) x] determination on a biochemical analyzer as compared to the POCT instrument. Also,no proportional error [y = 0.17 (0.07-0.27) + 1.06 (0.99-1.13) x] was found in the determination of Lac on a biochemical(CV = 1.79%) vs. POCT (CV = 4.37%) analyzer. Although a constant error was observed, it had no clinical significance(mean BIAS = -17.4%; calculated RCV = +/-75.9%)

CONCLUSIONS

In conclusion, repeating the critical values of K, Na, Cl, Glc, and Lac on the POCT rather than on the biochemical analyzeris justified because the analytical process for determination on the POCT analyzer takes much less time and the criticalresults are reported to the physician much earlier.


W175


COMPARATIVE STUDY OF THE DETERMINATION OF BIOCHEMICAL TESTS IN THE COBAS 8000 AND POCT I-STATANALYZERS

M. Zarate 1, M. Giménez Blanco 1, R. Rubio Sánchez 1, M.d.C. Esteban De Celis 1, M.d.M. Viloria Peñas 11Hospital Universitario de Valme

BACKGROUND-AIM

In recent years, the implementation of Point-of-Care Testing (POCT) equipment has increased considerably, especiallyblood gas analyzers. However, new trends suggest that these same analyzers can perform more comprehensive tests,including basic biochemistry (glucose urea, creatinine, sodium and potassium), becoming tools of high diagnostic valuein few minutes.The aim of this study has been to evaluate the correlation and transferability of the results between the COBAS® 8000analizer (Roche Diagnostics) and the POCT i-STAT® equipment (Abbott) for measured parameters relative to the basicbiochemistry profile.

METHODS

60 samples from our hospital’s Emercency Room were included to determine a basic biochemistry profile. The analysisof the samples was carried out in both equipment following a correlative order to minimize preanalytical error: first inthe COBAS® 8000 analyzer and followed on the POCT i-STAT®.Statistical analysis for method comparison was performed using Passing-Bablok regression and the analysis of themean differences using the Bland-Altman method with the program Statistical MedCalc v18.2.1.

RESULTS

For glucose with a 95% IC of slope 0,8182 to 1,2683, Pearson correlation coefficient (r) was 0,803 (p=0,0002); urea witha 95% IC of slope 0,5000 to 1,0000, r was 0,937 (p<0,0001); creatinine with a 95% IC of slope 0,6897 to 1,3043, r was0,915 (p<0,0001); sodium with a 95% IC of slope 1,0000 a 2,5000, r was 0,665 (p=0,0049) and potassium with a 95%IC of slope 0,4000 a 1,2000, r was 0,867 (p<0,0001).

CONCLUSIONS

Results obtained on the POCT i-STAT® analyzer are traceable and comparable to those of the COBAS® 8000 analyzeras there are neither systematic or proportional differences in any of the outcomes, so it is possible to ensure thetransferability of the results between both equipment.The correlation coefficient shows a strong positive association between both analyzers. Since the POCT i-STAT®analyzer is smaller, it is better for handling and transportation. In addition, the speed to give a result implies a reductionin the time for taking decisions and therapeutic intervention.


Date post:	19-Feb-2022
Category:	Documents
Upload:	others
View:	4 times
Download:	0 times

Point of Care testing, critical care, emergency medicine ...

Documents